CS5228 Knowledge Discovery And Data Mining - Project

Introduction

This is a solution to the CS5228 Knowledge Discovery and Data Mining Project competition on Kaggle (https://www.kaggle.com/c/1910-cs5228-knowledge-discovery-and-data-mining/overview).

Prerequisites

• Python 3.6 or above
• Virtualenv
• Kaggle API key set up (https://github.com/Kaggle/kaggle-api)

Directory Structure

|- src | |- ml_feat.py : Code for LightGBM based approach | |- nn.py : Code for Neural Network based training | |- ensemble.py : Code to combine outputs of both | |- analyze.ipynb: Basic data analysis |- requirements.txt : Python requirements file |- README.md : This file |- outputs : Folder containing generated CSV files

Installation and Deployment Instructions

• Run run.sh script on your machine.
• The final labels file will be generated as output.csv in the outputs directory